INRIA a CCSD electronic archive server
Not a member yet
122212 research outputs found
Sort by
Mopsa-C with Trace Partitioning and Autosuggestions (Competition Contribution)
No full textInternational audienceWe present advances we brought to Mopsa for SV-Comp 2025. Most notably, Mopsa now supports bounded trace partitioning, constant widening with thresholds, and can check that all memory has been correctly deallocated. Further, Mopsa now integrates a sound support of bitfields. While Mopsa at SV-Comp previously relied on a fixed, homogeneous set of configurations to verify tasks, it can now automatically leverage semantic information from a previous analysis to trigger heuristic precision improvements in further analyses. With these improvements, Mopsa wins a silver medal in the SoftwareSystems category and ranks fifth in the NoOverflows category
Are LSTM and conceptual rainfall-runoff models able to cope with limited training datasets under diverse hydrometeorological conditions?
No full textInternational audienceAs climate change exacerbates variability and non-stationarity in rainfall patterns, it is crucial to assess the predictive capabilities of forecasting models. Previous researches on rainfall-runoff modeling have focused on the impact of training dataset size on Artificial Neural Networks (ANNs) results, with limited consideration of hydrometeorological diversity. This study first evaluates the influence of the training dataset length (1 to 15 years) on performance of a Long Short-Term Memory (LSTM) and a traditional conceptual model, Superflex, across 10 validation years. Next, training years are categorized based on hydrometeorological diversity (wetter, standard, drier). This clustering allows for experiments where models are trained on data from similar or different clusters, enhancing understanding of how data diversity, and therefore climate change, can affect model performance. Results indicate that the LSTM model is highly sensitive to training length, showing poor performance with short datasets (below three years), reaches similar performance to Superflex around six training years on average, and overperforms with 15 years of training. Conversely, Superflex maintains rather constant performance levels regardless of the dataset length. LSTM model benefits from diverse 1 training data, achieving higher accuracy and reliability when trained on years with diverse hydrological typology. Despite their potential to outperform traditional models (with six or more training years on average), LSTM models are highly dependent on the quality and diversity of training data. In climate change scenarios, caution is needed when applying LSTM models to unfamiliar conditions, as their predictive accuracy may decline more rapidly than that of more traditional hydrological models
Deterministic limit of a stochastic individual-based model of biological tissue interacting with diffusing chemicals
No full textMathematical models of biological tissues are a promising tool for multi-scale data integration, computational experiments and system biology approaches. While some data and insights originate at the cellular level, macroscopic mechanisms emerge and are observed at the tissue scale, making tissue modeling an inherently multi-scale process. Consequently, tissue models can be broadly categorized into individual-based or continuous population-based. In this paper, we first introduce a generic individual-based model of biological tissue including the main regulation processes such as cell division, differentiation, migration and death, along with cell-cell mechanical interactions. This cell population interacts with diffusing molecules, which are consumed or produced by cells and, in turn, regulate cellular behavior. This model is a measure-valued, piecewise-deterministic Markov process, coupled with reaction-diffusion PDEs. We establish the well-posedness of the model and rigorously derive its large-population deterministic limit
Relax and penalize: a new bilevel approach to mixed-binary hyperparameter optimization
No full textInternational audienceIn recent years, bilevel approaches have become very popular to efficiently estimate high-dimensional hyperparameters of machine learning models. However, to date, binary parameters are handled by continuous relaxation and rounding strategies, which could lead to inconsistent solutions. In this context, we tackle the challenging optimization of mixed-binary hyperparameters by resorting to an equivalent continuous bilevel reformulation based on an appropriate penalty term. We propose an algorithmic framework that, under suitable assumptions, is guaranteed to provide mixed-binary solutions. Moreover, the generality of the method allows to safely use existing continuous bilevel solvers within the proposed framework. We evaluate the performance of our approach for two specific machine learning problems, i.e., the estimation of the group-sparsity structure in regression problems and the data distillation problem. The reported results show that our method is competitive with state-of-the-art approaches based on relaxation and roundin
Schrödinger evolution on surfaces in 3D contact sub-Riemannian manifolds
No full textLet M be a 3-dimensional contact sub-Riemannian manifold and S a surface embedded in M . Such a surface inherits a field of directions that becomes singular at characteristic points. The integral curves of such field define a characteristic foliation F . In this paper we study the Schrödinger evolution of a particle constrained on F . In particular, we relate the self-adjointness of the Schrödinger operator with a geometric invariant of the foliation. We then classify a special family of its self-adjoint extensions: those that yield disjoint dynamics
Stochastic Tangential Pareto Dynamics Provably Samples the Whole Pareto Set
No full textThe framework of stochastic multi-objective programming allows for the inclusion of uncertainties in multi-objective optimization problems at the cost of transforming the set of objectives into a set of expectations of random quantities. The stochastic multigradient descent algorithm (SMGDA) gives a solution to these types of problems using only noisy gradient information. However, a bias in the algorithm causes it to converge to only a subset of the whole Pareto front, limiting its use. We analyze the source of this bias and prove the convergence of SMGDA to a stationary point in the nonconvex L-lipschitz smooth case. First, based on this analysis, we propose to reduce the bias of the stochastic multi-gradient calculation using an exponential smoothing technique. We then propose a novel approach to exploring the whole Pareto set by combining the debiased stochastic multigradient with an additive non-vanishing noise that guides the dynamics of the iterates tangential to the Pareto set. We finish by proving that our algorithm, Stochastic Tangential Pareto Dynamics (STPD), generates samples concentrated on the whole Pareto set
SPIRIT-DEFINE explanation and elaboration: recommendations for enhancing quality and impact of early phase dose-finding clinical trials protocols
No full textInternational audienceTransparent and accurate reporting in early phase dose-finding (EPDF) clinical trials is crucial for informing subsequent larger trials. The SPIRIT statement, designed for trial protocol content, does not adequately cover the distinctive features of EPDF trials. Recent findings indicate that the protocol contents in past EPDF trials frequently lacked completeness and clarity. To address this gap, the international consensus-driven SPIRIT-DEFINE checklist was developed through a robust methodological framework for guideline development, with the aim to improve completeness and clarity in EPDF trial protocols. The checklist builds on the SPIRIT statement, adding 17 new items and modifying 15 existing ones.The SPIRIT-DEFINE explanation and elaboration (E&E) document provides comprehensive information to enhance understanding and usability of the SPIRIT-DEFINE checklist when writing an EPDF trial protocol. Each new or modified checklist item is accompanied by a detailed description, its rationale with supportive evidence, and examples of good reporting curated from EPDF trial protocols covering a range of therapeutic areas and interventions. We recommend utilising this paper alongside the SPIRIT statement, and any relevant extensions, to enhance the development and review of EPDF trial protocols.By facilitating adoption of the SPIRIT-DEFINE statement for EPDF trials, this E&E document can promote enhancement of methodological rigour, patient safety, transparency, and facilitate the generation of high-quality, reproducible evidence that will strengthen the foundation of early phase research and ultimately improve patient outcomes
Assessing the effects of manual therapy on pain in patients living with persistent non-specific low back pain (PNSLBP): Which evaluation criteria and measurement tools are used in randomised controlled clinical trials? A systematic review
No full textInternational audienceObjective: To identify the evaluation criteria and measurement tools that are used in Randomised Controlled Trials (RCT) to assess the effects of manual therapy on pain in adults living with PNSLBP.Methods: RCT were included if: participants were adults (18–65 years old) with PNSLBP, experimental group received manual therapies (osteopathy, physiotherapy and chiropractic) and they had been published in English or French since 2010. The search was conducted between May 2021 and April 2023, using the Cochrane Library, Ebscohost, EMBASE, MEDLINE Pubmed, PEDro, ScienceDirect and Scopus databases. Three independent reviewers have checked eligibility. The PEDro scale have been used for quality appraisal.Results: In the 29 studies included, 131 measurement tools were identified in three main areas: 76 % of the measurement tools were related to life impact (which 24 % related to pain intensity and 23 % to functional difficulties) and 24 % were related to pathophysiological manifestations (which 8 % related to Range of motion).Conclusion: Most measurement tools focus on pain intensity and physical functioning using scales and questionnaires. Two perspectives must be considered: responding to current recommendations by standardizing the measurement tools relating to life impact and physiopathological manifestations, and using biomechanical markers making it possible to evaluate patients in daily life situations
A characterization of inner product spaces via norming vectors
No full textInternational audienceA finite-dimensional normed space is an inner product space if and only if the set of norming vectors of any endomorphism is a linear subspace. This theorem was proved by Sain and Paul for real scalars. In this paper, we give a different proof which also extends to the case of complex scalars
